[R-sig-ME] Article on using observation-level random effects vs beta-binomial models for overdispersed binomial data

Viechtbauer Wolfgang (STAT) wolfgang.viechtbauer at maastrichtuniversity.nl
Tue Jul 21 21:04:09 CEST 2015

Dear All,

An article was just published in PeerJ that may be of interest to some readers of this mailing list:

Harrison, X. A. (2015). A comparison of observation-level random effect and beta-binomial models for modelling overdispersion in binomial data in ecology & evolution. PeerJ 3:e1114 https://dx.doi.org/10.7717/peerj.1114 

lme4 was used for fitting the model with observation-level random effects.

One thing that I would have liked to see is a discussion of how the choice of estimation method (i.e., PQL, Laplace, (A)GQ) may impact the accuracy of the methods. Lots of previous research on that (Austin, 2010; Kim et al., 2013; Li et al., 2011; Zhang et al., 2011), indicating that this can make quite a bit of a difference, especially when the number of clusters is low (which was in fact the focus of this article).


Wolfgang Viechtbauer, Ph.D., Statistician | Department of Psychiatry and    
Neuropsychology | Maastricht University | P.O. Box 616 (VIJV1) | 6200 MD    
Maastricht, The Netherlands | +31 (43) 388-4170 | http://www.wvbauer.com     

More information about the R-sig-mixed-models mailing list